"++remote-shell",
"ssh",
"++nodelist",
"/root/machinefile",
"+p2",
"/usr/local/NAMD_2.7b1_Linux-x86/namd2",
"/root/run/NPT.conf",
]
# jd.working_directory = "/root/run/"
jd.output = "stdout.txt"
jd.error = "stderr.txt"
jobs = []
for i in range(0, NUMBER_JOBS):
print "Start job no.: " + str(i)
sj = bigjob_cloud.subjob(bigjob=bj)
sj.submit_job(jd)
jobs.append(sj)
# busy wait for completion
while 1:
try:
number_done = 0
for i in jobs:
state = str(i.get_state())
print "job: " + str(i) + " state: " + str(state)
if state == "Failed" or state == "Done":
number_done = number_done + 1
if number_done == len(jobs):
break
# !!Adjust!!
jd.working_directory = "/home/luckow/run/"
jd.output = "stdout_tg.txt"
jd.error = "stderr_tg.txt"
print "**************** Start SubJob Submission ************** "
subjob_start = time.time()
jobs_ec2 = []
jobs_nimbus = []
jobs_tg = []
jobs_condor = []
number_started_jobs= 0
total_number_jobs = NUMBER_JOBS_NIMBUS+NUMBER_JOBS_EC2+NUMBER_JOBS_GRID+NUMBER_JOBS_CONDOR
for i in range(0, NUMBER_JOBS_NIMBUS):
print "Start job: " +str(number_started_jobs + 1) + " on Nimbus"
sj_nimbus = bigjob_cloud.subjob(bigjob=bj_nimbus)
jd_nimbus.output = "stdout_nimbus.txt."+str(number_started_jobs+1)
jd_nimbus.error = "stderr_nimbus.txt."+str(number_started_jobs+1)
sj_nimbus.submit_job(jd_nimbus)
jobs_nimbus.append(sj_nimbus)
number_started_jobs = number_started_jobs + 1
for i in range(0, NUMBER_JOBS_EC2):
print "Start job no.: " +str(number_started_jobs + 1) + " on EC2"
sj_ec2 = bigjob_cloud.subjob(bigjob=bj_ec2)
jd_ec2.output = "stdout_ec2.txt."+str(number_started_jobs+1)
jd_ec2.error = "stderr_ec2.txt."+str(number_started_jobs+1)
sj_ec2.submit_job(jd_ec2)
jobs_ec2.append(sj_ec2)
number_started_jobs = number_started_jobs + 1
cloud_bigjobs.append(bj_cloud)
# reset intervall timer
intervall_timer=time.time()
print "Pilot Job/BigJob URL: " + bj_tg.pilot_url + " State: " + str(bj_tg.get_state_detail()) + " Time since launch: " + str(time.time()-start)
for i in cloud_bigjobs:
print "Pilot Job/BigJob URL: " + i.pilot_url + " State: " + str(i.get_state_detail()) + " Time since launch: " + str(time.time()-start)
if number_started_jobs < NUMBER_JOBS:
for bj in cloud_bigjobs:
if bj!=None and str(bj.get_state_detail())=="Running":
print "Cloud: Free nodes: " + str(bj.get_free_nodes())
if int(bj.get_free_nodes()) >= int(jd_nimbus.number_of_processes):
print " Start job: " +str(number_started_jobs + 1) + " on " + str(jd_nimbus.number_of_processes)
sj_nimbus = bigjob_cloud.subjob(bigjob=bj)
jd_nimbus.output = "stdout_nimbus.txt."+str(number_started_jobs+1)
jd_nimbus.error = "stderr_nimbus.txt."+str(number_started_jobs+1)
sj_nimbus.submit_job(jd_nimbus)
jobs_nimbus.append(sj_nimbus)
number_started_jobs = number_started_jobs + 1
if str(bj_tg.get_state_detail())=="Running":
print "TG: Free nodes: " + str(bj_tg.get_free_nodes())
if int(bj_tg.get_free_nodes()) >= int(jd.number_of_processes):
print "Start job no.: " +str(number_started_jobs + 1)
sj_tg = bigjob.subjob(advert_host)
jd.output = "stdout_tg.txt."+str(number_started_jobs+1)
jd.error = "stderr_tg.txt."+str(number_started_jobs+1)
sj_tg.submit_job(bj_tg.pilot_url, jd)
jobs_tg.append(sj_tg)
